Page 44 - Cardiac abnormalities after aneurysmal subarachnoid hemorrhage

Chapter 3
Assessment of outcome: Delayed cerebral ischemia:
The study coordinator (IB) entered all episodes of clinical deterioration in an electronic database. Following this procedure, two neurologists (FK,MJ), who had access to all clinical data except the cardiac data, assessed whether these clinical deteriorations fulfilled the criteria for probable or definite DCI. Probable DCI was defined as a focal deficit or deterioration of level of consciousness not explained by rebleeding, hydrocephalus, complication from aneurysm treatment, infection, or metabolic disturbances, but without new hypodensities on a repeated CT scan. Infection was defined as fever, leucocytosis, and increased C-reactive protein, or positive cultures along with clinical signs of infection for which antibiotic treatment was started. Definite DCI was defined as development of focal neurological signs or deterioration of the level of consciousness, or both, with evidence of cerebral infarction on CT-scan or at autopsy. Any new hypodensity on the CT scan without an obvious explanation such as neurosurgical or endovascular intervention, or perihematomal edema were scored as definite DCI even in the absence of clinical symptoms. Probable and definite DCI were taken together as one event.
Poor outcome:
Patients were contacted by phone or in the outpatient clinic three months after admission and outcome was scored according to the Modified Rankin Scale (MRS).11 Poor outcome was defined as a MRS score > 3.
Statistical analysis:
Baseline characteristics were analyzed using descriptive statistics. Univariable risk ratios (RR) for baseline determinants and outcomes at three months (DCI, death and poor outcome) with corresponding 95% CI were calculated with Poisson regression with robust standard errors.12, 13 For determinant-outcome combinations that yielded a significance level of 5%, multivariable adjustments were performed with Poisson regression. Adjustments are reported for those factors that changed the crude risk ratio by more than 5%. Missing values were excluded by casewise deletion. Hijdra score was dichotomized on the median, WMSI was dichotomized on one standard deviation above the median.
Additionally, we made Kaplan-Meier curves for cumulative percentages of death according to presence or absence of WMAs. Because rebleeding may increase the risk of DCI and poor outcome the chance of cardiac abnormalities, we performed a sensitivity analysis with exclusion of patients who had an episode of rebleeding after
42